Distributed simulation-based analytics approach for enhancing safety management systems in industrial construction

Although methods for assessing and simulating the influence of safety-related measures on safety performance have been proposed, practical applications remain limited. Data required by these methods are dispersed across departments, necessitating the development or redesign of data warehouses. This research proposes a simulation-based analytics approach to enhance safety management system (SMS) decision making using distributed simulation to overcome limitations associated with previous approaches. This distributed simulation approach is used to (1) integrate historical data without modifying data-warehouse structures (i.e., data fusion component), (2) link data to an artificial neural network-based analysis component for determining the influence of safety-related measures on incident levels, (3) connect data and analysis components to existing simulation components, and (4) combine the outputs, resulting in a comprehensive safety performance evaluation system to examine incident levels. Results demonstrate that this approach successfully fuses and integrates data from several sources with analysis and simulation components in a cost-, labor-, and time-efficient manner. A distributed simulation-based analytics approach represents a considerable opportunity for industrial construction companies to more effectively use historical data, analysis tools, and simulation models.